54
Chapter III
Carotenoid-based and depigmented patches on Narcissus Flycatcher
55
degrees of condition dependence (Badyaev et al. 2001, Leskinen et al. 2012, Guindre-Parker and Love 2013). Moreover, the advantages and developmental trade-offs may change with age (Boves et al. 2014, Grunst et al. 2014) and environment (Alatalo et al. 1986, Lifjeld and Slagsvold 1988, Taff et al. 2013). Thus, knowledge of developmental and functional interrelationships among the components of a condition-dependent trait is essential for understanding the evolutionary changes in the trait (Endler 1992, Badyaev and Hill 2001, Guindre-Parker and Love 2013).
Plumage colours are developed during specific moulting periods and hence reflect the nutritional status at the time of the moult (Hill and McGraw 2006, de la Hera et al. 2010. Grindstaff et al. 2012). Moulting seasonality differs among bird species and populations (Ginn and Melville 1983, Svensson 1992, Barta et al. 2008, Newton 2009). Some migratory birds have two moulting periods. One is a partial pre-breeding moult in the wintering area, and the other is a complete post-breeding moult in the breeding area (Ginn and Melville 1983, Barta et al. 2008, Newton 2009).
Breeding plumages of these species bear patches of feathers constructed in the wintering area and some produced in the breeding area (Svensson 1992). Therefore, plumages that develop in different areas may indicate the condition of individuals during different periods. If an individual’s condition
is stable, and there is a correlation between plumage colourations of the pre-breeding moult and those of the post-breeding moult, the two different plumages can be work as redundant signals to convey the condition of the individual throughout the year. However, if there is no correlation between these two plumage colourations, the two different plumages can show different aspects of male quality. In addition, the moulting seasonality of some migratory birds differs between mature and immature. For example, yearlings should wear mature-male-like plumage to bear sexual traits in the partial pre-breeding moult; however, mature male does not have to moult excessively to show
56
this. Thus, even if the breeding moults of a yearling and mature male look similar, the basic message of the plumages might be different. It is crucial to determine the correlations among multiple morphological traits within age classes in order to reveal the role of multiple traits in birds.
The Narcissus Flycatcher Ficedula narcissina is a sexually dimorphic, small, migratory passerine, breeding in Northeast Asia and wintering in Southeast Asia (del Hoyo et al. 2006, Töpfer 2006). It is socially monogamous; nests in wood cavities and both parents care for nestlings and fledglings (Nakamura and Nakamura 1995, Okahisa et al. 2012). The male Narcissus Flycatcher has bright yellow carotenoid-based patches on the breast and rump, and depigmented wing patches on the wings (mainly on the inner greater and medium coverts, and in some cases also on the tertials).
In addition, the summer plumage colourations of yearling males differ significantly from those of mature males. The entire upper part of a mature male is jet black. On the other hand, the primaries, secondaries, tertials, greater coverts, medium coverts, lesser coverts, and alula of yearling males have some female-like, brown juvenile feathers, and the number of brown feathers they have varies (Yamashina 1941, Okahisa et al. 2013). First-winter males moult body feathers (including forehead, breast, rump and all under parts), coverts, tertials (including inner secondaries in a few cases), and tails (all or partial). They bear sexual traits (black upper parts, yellow breast and rump patches, and white wing patches) during the pre-breeding moult in the wintering area. In contrast, mature males have a complete post-breeding moult in the breeding area, mainly in July or August. They also have a partial pre-breeding moult of some body feathers (including the breast), a few coverts, and the tail feathers (all or partial) in the wintering area (Yamashina 1941). In addition, older individuals have a more reddish iris colour, and darker and denser tibial feathers (Okahisa et al. 2011).
57
In this study, the carotenoid-yellow patches on the breast and the rump, and the depigmented white patches of male Narcissus Flycatchers were focused. Males display their breast plumage to females to attract them for pairing by raising their heads and shaking them slowly and smoothly. In addition, males flush their plumage colourations in display flights to attract mates. They also strongly attack the body plumage in male-male competitions. Thus, these plumage colourations may play an important role in the sexual selection. Carotenoid-based colourations are recognized as an important signal, yet the essential mechanisms that maintain their honesty is not well understood (Hadfield and Owens 2006). Animals can only obtain carotenoids from their diet (Goodwin 1984).
The amount deposited in ornaments reflect the individual’s ability to acquire and assimilate these pigments (Peters et al. 2008). In addition, the message of the depigmented white patches of the Ficedula flycatcher is not well understood (see Török 2003). It can weakly reflect the prior condition of individuals (Török 2003) but is usually treated as a sexual trait indicating genetic quality (Potti 1993, Sheldon et al. 1997, Sheldon and Ellegren 1999, Sanz 2001, Sirkiä and Laaksonen 2009, de Heij et al. 2011, Laczi et al. 2011). To reveal the benefits and developmental trade-offs of such multiple traits, the degree of conditional dependence, while considering both age and environment were tested.
The aims of this study are to (1) reveal the changes in the colouration of the carotenoid patches and the size of the depigmented white patches with age, (2) determine correlations between multiple traits, (3) examine the effects of territory quality on the colours of the plumage that develop in post-breeding moulting, and (4) determine whether these multiple traits reveal multiple messages conveying different condition dependencies or redundant signals.
58 METHODS
Field research
Data were collected from 15 April to 15 June in 2010, 2011, 2012, and 2013, during the breeding season of the Narcissus Flycatcher, in Fuji Primitive Forest, Central Japan (35°27′N, 138°38′E; 60 ha; 1140 m above sea level). The study area consists of two forest types: deciduous broadleaf forest and evergreen coniferous forest. The deciduous broadleaf forests are dominated by old growth Japanese Oak (Quercus crispula), Siebold’s Beech (Fagus crenata), and Japanese Blue Beech (F.
japonica). The evergreen coniferous forests are dominated by old growth Japanese Cypress (Chamaecyparis obtusa) and Japanese Hemlock (Tsuga sieboldii) (Okahisa et al. 2012). Arthropods (Diptera, Lepidoptera, Hymenoptera, and Coleoptera) are more abundant in the broadleaf forest than the coniferous forest (Okahisa et al. in press), and the Narcissus Flycatcher in the broadleaf forest had larger clutch sizes than those in the coniferous forest (Yuji Okahisa submitting).
Each male was lured into a mist net using song playback within his territory or at a rain puddle. The age of each bird was determined by plumage characteristics (Okahisa et al. 2013). Males were uniquely ringed with a combination of aluminium (authorized by the Japan Environment Agency) and colour-coded rings. The following measurements were taken: body mass with a 30 g Pesola spring balance (to the nearest 0.1 g, Pesola AG, Rebmattli, Baar, Switzerland), wing length with Mitsutoyo sliding digital callipers (natural wing chord in 0.1 mm, Mitsutoyo Corp., Kawasaki, Kanagawa, Japan), and tarsus length with Mitsutoyo sliding digital callipers (to the nearest 0.1 mm).
The moulting status (feathers moulted in the wintering area or brown juvenile) of each greater covert, primary, secondary, tertiary, and rectrices were recorded. The sum of the moulted feathers on these parts was used as an indicator of the moulting status of an individual.
59
Singing males were searched in the study area every morning (03:30–11:30). When a singing male was observed, the male was followed and plotted the singing location on maps. The locations of singing points were plotted using the colour markings on trees every 20 m (made before the spring migration of Narcissus Flycatchers), and each marking was located with a GPS (Garmin, eTrex Legend, accuracy was ± 8 m). One observation trial was just 5 minutes, and the trial was repeated every day. Finally, the minimum convex polygon of their song locations was calculated with ArcGIS 10.1 (Environmental Systems Research Institute). It was considered as a flycatcher’s territory. To estimate the territory quality of the individual, the vegetation characteristics of the habitats were described in 50 randomly placed 10 m × 10 m quadrats in each forest type (deciduous broadleaf and coniferous evergreen). The number of each tree species and their diameter at breast height (DBH) values were recorded for each quadrat. Next, the area-weighted average proportion of broadleaf trees of each territory was calculated with ArcGIS. It is known that this proportion of broadleaf forest is a suitable explanatory factor for predicting prey abundance for the insectivorous birds in the study area (Okahisa et al. in press).
Colour analysis
Upon the capture, five breast feathers, five rump feathers, and three feathers of greater coverts were plucked from males for spectrometric analyses. The feathers were placed on the black background to follow the natural situation (Bentz 2013). The measurements of plumage colouration were obtained at the beginning of breeding. The spectral reflectance between 300 and 700 nm was measured using an Ocean Optics Jazz spectrometer (range 200–800 nm, Ocean Optics Inc., Dunedin, FL, USA), illuminated with both UV (deuterium tungsten-halogen bulb) and visible (tungsten-halogen bulb)
60
light sources, and a WS-1 white standard (Ocean Optics Inc., Dunedin, FL, USA). A bifurcated micron fibre optic probe was used at a 90° angle, 1 mm from the feather’s surface (Bentz 2013).
For the analysis of reflectance, I used AVICOL software (Gomez 2006). Initially, I summarized the reflectance data (Fig. 1). To compare the entire colour pattern of the bird, I also used the model of Endler and Mielke (Endler and Mielke 2005) following (Peters et al. 2008). For each reflectance plumage patch, I determined cone quantum catches based on spectral sensitivities of the four cones (VS: very short, S: short, M: medium, and L: long wavelength sensitive cones) used in the colour vision (Endler and Mielke 2005, Peter 2008). The sensitivity of the cone type was based on cone sensitivities for type U-eyes according to Appendix A in Endler and Mielke (2005). I then divided each cone quantum catch by the sum of all four and transformed these according to Kelber et al. (2003), to obtain three independent relative cone catches—x, y, and z—according to this transformation. Higher values of x represent greater stimulation of the L cone and lower stimulation of the M cone; higher y values represent greater stimulation of the S cone, and higher values of z greater stimulation of the VS cone (Endler and Mielke 2005, Peters et al. 2008). These relative cone catches can be represented as points in the tetrahedral avian visual space. I calculated the first principal component (PCxyz, Peters et al. 2008), which explained 78–94% of the each variation (Table 1). I also determined the white patch on the wing by scoring the size of the white patch on each feather from the greater coverts and tertials (Fig. 2). The greater coverts and tertials overlap each other and move easily. It makes calculation of the real total area difficult. Thus the sum of each wing patch size was used as an indicator of whole white patch size.
Statistical analysis
61
All statistical tests were carried out using R version 3.0.0. (R development core team 2013). To examine the difference in male morphology based on the age and study year, linear mixed models (LMMs) with Gaussian error and the identity link function were used. Linear models (LMs) were also used for testing these effects on wing patch reflectance. P-values for age and study years were obtained from a likelihood F-test. The male age used as a rank variable, but the study years were used as a factor variable. To test correlations among male characteristics and morphological repeatability of breast plumage colour, Pearson’s correlation test and Spearman’s rank correlation test were used. The morphological characteristics to the previous and present year were compared in the test of morphological repeatability. The LMMs with Gaussian error and the identity link function was also used for testing the effects of prior territory on morphological characteristics in the next year. In this case, the proportion of broadleaf tree was included as an explanatory variable. Study year and IDs of individuals were used as random factors. To test the effects of morphological change, characteristics in the year (t) – the trait in the year (t – 1) was used to determine the difference in a given morphological trait.
RESULTS
No differences in breast, rump, and wing patch reflectance among the age groups were found (Table 2), but the wing patch size significantly increased with age. The breast reflectance was different among the study years. The carotenoid-based pigmental breast and rump reflection were not correlated with each other in mature males or yearlings (Fig. 3, Pearson’s correlation test: breast and rump in mature male, t = 1.53, df = 69, P = 0.13; breast and rump in yearling, t = 0.96, df = 57, P = 0.34). In addition, no correlation was found between the depigmented wing patch reflections and the
62
patch size (Spearman’s rank correlation test: wing patch reflectance and wing patch size in mature
males S = 122.10, P = 0.95; wing patch reflectance and wing patch size in yearling S = 21.19, P = 0.50).
No correlations were found between other morphological traits (Pearson’s correlation test:
wing patch reflectance and breast in mature males t = -0.21, d = 9, P = 0.84; in yearlings, t = 0.14, df
= 6, P = 0.89; wing patch reflectance and rump in mature males, t = 1.25, df = 8, P = 0.25; in yearlings, t = 2.36, df = 6, P = 0.06; Spearman’s rank correlation test: wing patch size and breast plumage; yearlings, S = 18168.81, P = 0.93; mature males, S = 30190.76, P = 0.82; wing patch size and rump plumage; yearlings, S = 16165.05, P = 0.15; mature males, S = 23705.41, P = 0.34).
Wing patch size change had strong repeatability (Spearman’s rank correlation test, rho = 0.67, S = 592.05, P = 0.00097) but I found no other repeatability in plumage reflections (Pearson’s correlation test: breast reflection, t = 0.50, df = 15, P = 0.62; rump reflection, t = 0.41, df = 15, P = 0.69). Comparing the morphological characteristics in the year (t) and year (t + 1), the rump reflectance component of 2-year-old males significantly increased when they had occupied a territory with more broadleaf trees in the past year (Table 3, Fig. 4). This tendency was not found in mature males. The rump reflectance component indicates higher values of the L cone and lower stimulation of the M cone, greater stimulation of the S cone, and lower stimulation of the VS cone.
In other words, the rump plumage of 2-year-old males became coloured with the strongest UV reflection in broadleaf forest areas. While, the proportion of broadleaf tree in the territories did not solely explain white patch size on the flycatcher in the past year, but longitudinal change in wing patch size was significantly correlated with previous territory. The wing patches became large when individuals had occupied the territory in broadleaf dominant area (Table 4, Fig. 4). No effects of
63 territory on breast reflectance and its change were found.
DISCUSSION
Carotenoid-based colour indicates parental ability, foraging ability, and their own health (Hill and Montgomerie 1994, Linville and Breitwisch 1997, Olson and Owens 1998, Navara and Hill 2003), because birds must digest more carotenoids in order to bear blighter carotenoid-based ornamentations. The colouration of the rump feathers was correlated with territory quality in the previous year, but no correlation was found between age and other morphological characteristics. It suggests that the yellow colouration of the rump signal an individuals’ condition in the previous post-nuptial season. Interestingly, the rump colour in 2-year-old males was correlated with the previous territory quality. However, no correlation was found in old males. This result suggests that older males compensate for territory quality and obtain food equally well in the broadleaf forest and the coniferous forest. Grunst et al. (Grunst et al. 2013, see also Badaev and Duckworth 2003) argued that the association between the condition and carotenoid-ornamentation might decline with age because the carotenoid saturation with age masks the effect of condition on ornamentation (Copeland and Fedorka 2012). The results of rump plumage might support this argument. A correlation was found only in the first post-nuptial moulting of sub-adults. Thus, the rump feathers could be differently reliable in conveying individual condition in different age groups.
In contrast, the breast carotenoid-based pigmentation was correlated with neither age nor previous territory quality. This is probably because the Narcissus Flycatcher moults its breast feathers in the wintering area. If the social ranks in breeding area were correlated with those in
64
wintering area, the breast and rump plumage colouration should have been correlated. However, no correlation was found between breast and rump colouration. The results suggest that the carotenoid-based patches on the breast and rump have different messages. The breast colouration might indicate the condition in the wintering area while the rump colouration conveys condition in the breeding area.
I hypothesized that there may be different interrelationships among the multiple traits on a bird body because of the different moult seasonality of the yearlings and the mature males. The breast feathers and rump feathers of yearlings are moulted in the wintering area, and both types of feathers are carotenoid-pigmented. If both patches truly reflect the amount of carotenoid-containing food that they took in the wintering area, there should be a correlation; however, no correlations between the characteristics were found in yearlings, nor in mature males. These results suggest that even in the pre-breeding moult of a yearling male, the condition-dependences of the two carotenoid-based patches are different. The reason is not clear. It is possible that yearlings may have two moulting periods in the wintering area, or the mechanism of these two patches is different.
McGraw (2006) reviewed carotenoids in bird bodies and divided them into 24 types (i.e. lutein, zeaxanthins, β-carotene, canary xanthophyll, and astaxanthin). A bird uses seven types of carotenoids for its plumage at maximum. The types differ among patches on a single bird, as well as between the sexes and age groups in some bird species (Hudon et al. 1989, Stradi et al. 1998, McGraw 2006). When moulting birds were fed higher lutein and zeaxanthin, they only accumulated more zeaxanthin, but not lutein (MacGraw et al. 2004). The types of carotenoids that Narcissus Flycatchers accumulate in breast and rump feathers is not known. The spectrum is different (Fig. 1);
65
thus the two carotenoid patches on the flycatcher might respond differently to conditions because of differences in mechanism.
The white patch size of the Narcissus Flycatcher had considerable variation among individuals; however, the white patch size showed high repeatability within individuals. In addition, the patch was slightly larger in older male groups, similar to other Ficedula Flycatchers (Potti 1993, Török 2003, Laczi et al. 2011). Normally, the size of the depigmented patch of Ficedula flycatchers is treated as a sexual trait indicating genetic quality, with a larger white patch influencing extra-pair paternity and mating success (Potti 1993, Sheldon et al. 1997, Sheldon and Ellegren 1999, Sanz 2001, Sirkiä and Laaksonen 2009, de Heij et al. 2011) for increased genetic benefits, such as inheritance of the phenotype for future attractiveness and improved viability (de Heij et al. 2011, Laczi et al. 2011). The high repeatability of white patch size and enlargement with age could support the idea that depigmented patch size is genetically decided; however, a significant tendency was found for the patch size to increase in individuals living in broadleaf forests. This result is similar to the study of the Collared Flycatcher (Ficedula albicollis); the wing patch size of the Collared Flycatcher reflects the prior condition and signals viability (Török 2003). Combining these arguments and the results of this study, the wing patch size of the Narcissus Flycatcher might mainly indicate the genetic quality of males (i.e. viability), while the patch size may also slightly convey long-term condition. The condition-dependence and reliability of white pigmentation are still an ongoing issues (Török 2003, Guindre-Parker et al. 2013), because a larger white patch means a smaller melanin-based black part; thus the bearing cost would be smaller than others (Török 2003, Badyaev and Hill 2000). Here, I have shown the effect of the previous condition on the depigmented
66
white patch size. Further studies are required to reveal the mechanism and developmental trade-off of depigmented white patches.
There could be other reasons why the rump reflectance and the wing patch size of the flycatcher were correlated with the proportion of broadleaf tree in the prior territory because the plumage colouration was not measured immediately after post-breeding moulting, but rather during the following spring. In addition, I cannot rule out the possibility that there were individual differences in the amount of partial pre-breeding moult in mature males, similar to that of yearlings (Okahisa et al. 2012). If there were inter-individual differences in the amount of pre-breeding moult in mature males, the correlation between prior territory and plumage colour could be affected by this pre-breeding moult. For example, broadleaf-inhabiting dominant individuals arriving earlier in the wintering area, could spend a longer time undergoing their pre-breeding moult, and might thus be able to bear higher quality feathers (de la Hera et al. 2010, Dawson 2004). However, to our knowledge, such a carry-over effect of post-breeding condition on the winter moulting of birds has not been demonstrated. In addition, the reflectance of breast plumage that was moulted in the wintering area was not correlated with rump reflectance and white patch size. Thus, there might be no effect of individual differences in the amount of partial pre-breeding moult, or the effect may be limited.
The results of this study determined (1) differences in the condition-dependency of the breast, rump, and wing patches on male Narcissus Flycatchers, (2) there was no correlation among multiple traits, and (3) the effects of prior territory quality on plumage colours of feathers that develop during post-breeding moulting. These results have led to the conclusion that these multiple traits are multiple messages depending on different conditions. It is likely that the message of breast
67
feather colouration is from the wintering condition; the rump feather is from the previous breeding condition, and white patch size is mainly age-based (a sign of viability), although white patch size could also weakly signal previous condition in the breeding area. Theoretical models suggest that multiple condition-dependent signals as handicaps are evolutionarily unstable (Schluter and Price 1993, Johnstone 1996, Iwasa and Pomiankowski 1994), but the finding in the flycatcher opposes this.
Theoretical re-construction for the evolution of multiple conditional traits in different periods is needed.
68
Table 1. Principal Component of plumage colour index, x, y, z in the Endler & Mielke model
Breast Rump Greater Coverts
PC1 PC2 PC3 PC1 PC2 PC3 PC1 PC2 PC3
Standard deviation 1.53 0.77 0.23 1.59 0.65 0.21 1.68 0.43 0.10 Proportion of Variance 0.78 0.20 0.02 0.84 0.14 0.01 0.94 0.06 0.00 Cumulative Proportion 0.78 0.98 1.00 0.84 0.99 1.00 0.94 1.00 1.00 Factor loading (x) -0.98 0.08 0.18 -0.98 0.13 0.16 0.98 0.17 0.07 Factor loading (y) -0.88 0.46 -0.13 -0.93 0.35 -0.13 0.98 0.17 -0.07 Factor loading (z) 0.79 0.61 0.08 0.84 0.53 0.04 -0.94 0.35 0.00 Eigen value(x) -0.64 0.10 0.76 -0.61 0.20 0.76 0.59 0.39 0.71 Eigen value(y) -0.57 0.60 -0.56 -0.58 0.54 -0.61 0.59 0.40 -0.71 Eigen value(z) 0.51 0.79 0.33 0.53 0.82 0.21 -0.56 0.83 0.00
Table 2. Difference in patch colouration among age groups. F and p were calculated with LMM (random factor = IDs).
Yearling 2nd year 3rd or older Age Year
69
Traits Mean±SD n Mean±SD n Mean±SD n F p DF F p DF
Breast reflectance -0.01±1.53 64 -0.11±1.50 27 0.20±1.46 43 0.41 0.52 1, 132 2.40 0.07 3, 129
Rump reflectance -0.06±1.64 60 -0.03±1.25 24 0.12±1.60 41 0.27 0.60 1,123 1.18 0.32 3, 120
Wing patch reflectance*1 0.47±1.21 9 0.06±0.19 2 0.23±1.77 10 0.16 0.70 1, 19 6.09 0.02 1, 18
Wing patch size 2.90±0.64 55 2.95±0.31 24 3.17±0.51 48 5.00 0.03 1,125 0.29 0.83 3, 122
*1 When testing the difference in white path reflectance among age groups and years, only differences between yearling and individuals older than 3 years, year 2010 and 2012, were tested with LM
Table 3. Results of the F test with LMMs explaining the effect of prior territory on morphological characteristics in the present year Broadleaf tree in territory
2nd year old 3rd year or older
F P DF F P DF
Throat reflectance 0.10 0.76 1,13 0.37 0.57 1,13
Rump reflectance 11.25 0.004 1,14 0.42 0.53 1,13
White patch size 0.64 0.44 1,11 0.07 0.79 1,13
Wing chord 0.004 0.94 1,11 0.01 0.94 1,13
Tail length 0.29 0.60 1,11 0.29 0.60 1,12
70
Table 4. Results of the F test with LMMs explaining the effect of prior territory on changes in morphological characteristics Broadleaf tree in territory
F P DF
Throat reflectance change 0.006 0.94 1,12
Rump reflectance change 0.49 0.50 1,11
White patch size change 5.25 0.04 1,11
Wing chord 0.27 0.61 1,13
Tail length 0.003 0.96 1,12
71
Fig. 1. Reflectance spectra of plumage (reddish yellow breast patch, yellow back patch and white patch on greater coverts) colour of the Narcissus Flycatcher (Ficedula narcissina).
72
Fig. 2. Typical morphology of the wing patch on the greater coverts. These are scored (a)0, (b)1/4, (c)1/2, (d)3/4, (e)1.
73
-4 -2 0 2
-4-202
Breast reflectance
Rump reflectance
-4 -3 -2 -1 0 1
-4-202
Wing patch reflectance
Rump reflectance
Fig. 3. Correlation between the breast reflectance component and rump reflectance component, and wing patch reflectance and rump reflectance. Black dots indicate mature males (second summer plumage or older), and white dots indicate yearlings.
74
0.0 0.2 0.4 0.6 0.8 1.0
-2-1012
Proportion of broadleaf tree in previous territory
Rump reflectance
2nd year old 3rd year or older
0.0 0.2 0.4 0.6 0.8 1.0
-0.4-0.20.00.20.4
Proportion of broadleaf tree in previous territory
Wing patch size change
Fig. 4. (a) Correlation between the proportion of broadleaf trees in the territory in the past year and rump reflectance in the present year, and (b) between proportion of broadleaf trees in the territory in the previous year and wing patch size change.
75